dune

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly runs searches and fetches dataset schemas and query results from Dune (e.g., dune dataset search / dataset search-by-contract in references/dataset-discovery.md and dune query run in references/query-execution.md) including community-contributed/uploaded tables, and instructs the agent to parse those JSON responses (and use result.metadata.column_names per references/visualization-management.md) to construct SQL, visualizations, and follow-up actions—so untrusted, user-generated third‑party content directly informs tool use and next steps.